Dramatic structure is equally elusive and essential. If I had to narrow down my thinking about structure to three core principles, they would be:
Staging scenes in order to create a full feeling of invested anticipation and dramatic curiosity in the audience. As Alexander Mackendrick once put it: “Drama is anticipation mingled with uncertainty.”
Sequencing events so that the audience reacts emotionally to the things that do happen, ideally in a way that’s unique to this particular story.
Creating the conditions by which a script can deliver the goods with scenes or sequences that fulfill the implied social contract a script has with its audience concerning its chosen genre.
When I write scripts on my own, I rarely find myself utilizing the tried-and-true three act structure. Largely because its inherent emotional architecture is so well-worn.
In a typical three act feature script, you know you’re going to be introduced to a single main character, and then they’re going to face a dilemma, and then there’s going to be complications, and then there’s going to be a lowpoint, and then some kind of rally, etc.
Handled well, such a structure can be effective. But for me, the three act structure is rarely inspiring. And at this point of my career, I want to be inspired by what I write.
This is partially because I’ve been doing this awhile and I don’t want to fall into routine hackery. But it’s also because I have a directing and/or showrunning eye on all of my scripts. Which means, I want to write scripts that I would be inspired by enough to spend two years of my life making.
So perhaps, at this point, there isn’t enough uncertainty in the three act structure for me to utilize it when I’m writing an original script for myself. It’s not enough of a tight-rope walk to get me excited.
I like to have a little uncertainty in my writing. Uncertainty not only in regards to what’s going to happen to the characters, but also in a meta-sense: how is this all one story? how is this going to come together?
But it’s not just boredom that pushes me to look for alternatives to the linear three act structure. Sometimes, it’s just simply not the best approach for a story.
In my experience, the default Hollywood script these days is the story of a single main protagonist undergoing the inevitable rise and fall and re-rise of a three act structure. It’s so ingrained that I don’t think many writers, executives, producers, directors, etc really consider any alternatives to it at the present time, at least not in the development stage.
Small example: one film I’ve been interested in bringing back is the 1970s disaster pic like Airport or Poseidon Adventure or Earthquake. Part of what made these films so popular, in my opinion, was that they often presented a wide swath of ordinary people from different walks of life suddenly thrown together by some catastrophe.
A mechanic or retiree or janitor could end up proving being as essential to the fates of a group of people as some square-jawed man of authority. Often, these modest type of people would prove to be more essential.
It was a kind of democratic, ensemble storytelling approach that went back at least as far as John Ford’s Stagecoach from the 1930s, if not earlier. I think right now would be a great time to bring this sort of movie back, especially with how divided our culture has grown to be.
The plot would be: the shit hits the fan. Now, a whole bunch of people who would normally never talk to each other are suddenly reliant on each other to survive. Let’s see what happens.
To me, that’s good shit. But so far, when I bring up this disaster pic idea to industry people in 2023, every single person suggests shaping it into a Dwayne Johnson/Mark Wahlberg type of starring vehicle, where the story focuses on just one rugged individual who proves his public heroism while also balming whatever private wound ails him in act one (usually an estranged spouse or kid who he proves his worth to in act three).
Which is sort of the default template these days: one single rugged movie star lead character to put on the poster with some big explosion behind him, caught in a three act linear story that you know ends with him saving the day.
But what these old school disaster pics offered was something more unpredictable. And, to me, interesting. There’d be a dozen or more key characters. So you didn’t know who would survive, who would prove to be a coward, or who would end up a hero. As the story unfolded, you got to see the situation and the stakes involved from multiple different angles, with people of varying backgrounds and stations of life.
Some films since the 1970s have excelled at this. Independence Day. Armaggedon. To an extent, Tony Scott’s underrated final film Unstoppable.
But by and large, we don’t tell make these kinds of movies anymore.
I think storytelling has gotten a lot more conservative and risk averse and predictable in the last 50 years. Star vehicles. Three act structures. Recycled IPs. Recently added to the mix: true-ish stories about the rise of some random corporate brand, aka the inspiring untold story of Spam or whatever.
In most theaters, audiences are being told variations of the same four or five stories. So, how to get outside of these boring boxes?
One alternative is telling a story in a non-linear fashion.
I’ve seen people criticize non-linear storytelling with some variants of “it’s just being clever” or “if it was a good story, you wouldn’t need to tell it out of order” or “it looks like someone’s copying Tarantino.”
But the non-linear approach to telling a story has been around for a long, long time. Homer would jump into his epics in the middle of the plot, then fill in exposition as he went along. Hamlet starts with his father already dead and we only find out what happened to him later.
For my generation, Tarantino became so identified with non-linear storytelling largely, I think, because that mode of cinematic structure had been largely abandoned by mainstream Hollywood movies in the 1980s because studios — does this sound familiar? — thought they’d figured out how to safely and reliably repackage variations on the same easily digestible four or five stories over and over again.
I mean, okay, you would see arthouse non-linear outliers like Jim Jarmusch’s Mystery Train, which would interweave stories that occured in the same story universe but which were also largely self-contained (like in Pulp Fiction). But more often than not, in the 1980s linear storytelling had gotten to be interchangeable with accessible storytelling, at least until Pulp Fiction blew things up for awhile.
But this wasn’t always the case. A shit ton of film noir and noir-influenced movies in its post-war heyday offered non-linear narratives. Sunset Blvd. opens up with William Holden already dead, then we find out how we get there. The 1946 version of The Killers opens with Burt Lancaster getting murdered, then we find out why. Stanley Kubrick’s The Killing forensically examines a racetrack robbery through various timelines and perspectives so we can find out exactly how things went so wrong.
These are all artful, fantastic films. But they were also aimed at general audiences. They weren’t arthouse outliers. They were entertainment. The writers and directors and editors simply decided these particular stories were more gripping to mainstream audiences when told in a non-linear order.
You can go back to Citizen Kane, which starts with the main character’s death and his famous last words. Next, we get a newsreel overview of his life. And with these elements established, we now begin to follow a newsman as he moves around freely in time, interviewing different associates from Kane’s life in order to understand him and why he died uttering the word “Rosebud.”
Pitch this kind of structure to a studio today and they’ll just say “well told” and you won’t hear back from them again.
Lawrence of Arabia likewise begins with Lawrence dying. Raging Bull doesn’t start with Jake LaMotta’s death, but it does start with him fat and nearly unrecognizable, practicing a monologue in his dressing room. With this preamble, the film goes back in time to show us how he got there.
If a story needed to move around in time in order to be told, both storytellers and audiences seemed more accepting of it than they do now.
Okay, so I prefer old movies to new ones. Not a surprise if you know me.
But my main point is: with all of these old movie examples, this temporal freedom allows the films to: 1) grab our attention and curiosity right away, 2) posit a very specific over-arching dramatic question for us to ponder as we experience the story.
By starting off these films with the end of these character’s lives, the storytellers sacrifice suspense (what’s going to happen?) for intrigue and resonance (how did this happen? and why does this matter?).
Some filmmakers will still move around in time freely. Tarantino, obviously. Nolan. Wes Anderson. Scorsese. James Gunn. Guy Ritchie’s Wrath of Man was kind of a blast. But by and large, this narrative freedom seems to be the purview of pretty established auteur dude types. But in the world of traditional studio feature development — where largely anonymous writers like me pitch on properties and get hired and fired and replaced — I’m not seeing a whole lot of temporal playfulness going on.
In the world of features, it seems moving around in time has become a bit of a lost art. Less so in TV. If anything, the flash-forward framing device has gotten a bit cliche, probably because it proved so effective in the Breaking Bad pilot.
You can name dozens of examples like this. The whole formula of weird opening scene followed by a two weeks earlier chyron card at the start of pilot or episode has become that, a formula.
And yet, damn if it isn’t effective to start a White Lotus season with the image of a coffin at an airport, or of a swimmer at a fancy resort discovering a dead body, and then going backwards in time to tell us how we got there.
With that kind of non-linear framing device within the murder mystery genre, now every scene going forward is infused with dramatic questions: is this the person who ends up dead? is this the person who killed them? is this seemingly everyday encounter going to be the motivation for the crime?
The genre-specific framing device Mike White has devised for his show carries so much inherent dramatic suspense in it that it allows him plenty of real estate to explore his characters and his themes with more specificity and depth than the genre usually offers.
In a lot of TV scripts, the flash-forward reads like a gimmick. But it’s not a gimmick in White Lotus. Why? Because it’s a fantastic way of framing his entire story within the expectations of its genre, grounding his scenes from the get-go in that genre’s key dramatic questions: who was killed? who killed them? why? and how will this effect everyone else?
If anything, a show like White Lotus picks up the baton from movies like The Poseidon Adventure. Instead of a disaster, it’s a murder at a high class resort that pulls these wildly different people into a single story. But there’s the same uncertainty and the same intermingling of wildly different types of people in both the new school murder mystery show and the old school disaster pic.
But back to non-linearity. Described in a certain light, Reservoir Dogs sounds intolerably clever and postmodern: a diamond store heist film where you never actually see the heist take place.
But like so many filmmakers that I love, Tarantino uses a seemingly lowbrow, populist genre in order to explore very specific character dynamics and themes. First scene, he drops us into a diner conversation where we meet all of the characters at once in a dizzying swirl of words and tough-guy codes. Then the title drops and we jump right into the aftermath of a heist-gone-wrong.
It’s a ballsy move. But as I continue watching Reservoir Dogs thirty years after it first blew my teenage mind, I’ve come to understand that cutting away the obligatory heist — and the usual narrative shoe-leather involved in staging it — is what frees Tarantino to deliver genre-specific goods that he otherwise wouldn’t have been able to offer.
In this initially disorienting jump from the talky diner scene to the primal scene of Mr. White (Harvey fucking Keitel) and Mr. Orange (Tim fucking Roth) driving away in a car with Orange bleeding and screaming like a gutted animal, Tarantino immediately stages two of the three main dramatic questions the film will answer:
How did this heist go wrong?
Who the fuck are these guys?
Then within a matter of minutes, Mr. Pink (Steve fucking Buscemi) arrives at the same abandoned warehouse meeting spot where White and Orange are at. And he immediately begins explicitly asking the third main dramatic question:
Who is the rat?
Once you land on these three questions as being the primary dramatic questions that QT’s story is designed to answer, you realize that you don’t actually need to show the heist to answer them.
Sure, you could show Mr. Blonde (Michael fucking Madsen) going on a kill-crazy ramage in the jewelry store. But doing that would rob the film’s masterful move of having Pink and White talk up about how crazy Blonde’s rampage was, and letting the viewer imagine it, then having Pink and White go after each other, only to have Mr. Blonde himself suddenly show up with his cavalier psychopathy.
By telling the story out of order, Tarantino answers a key dramatic question (“How did this heist go wrong?”) in a much more satisfying, more entertaining manner than simply playing things out in order.
Likewise, a linear version of Reservoir Dogs would probably start with Mr. Orange getting an undercover assignment to infiltrate this heist crew. And we would be suitably tense as he tries to learn his undercover anecdote and tries to convince the crew that he’s one of them.
In this linear version, the dramatic questions would be more like the dramatic questions in most undercover cop movies: will the cop get caught? will he bring down the crew? will he begin to feel his loyalties shift?
But I’ve seen that movie before, and I think who is the rat? is a much more fun question to answer. And it’s one that can probably only be answered well in a non-linear fashion, through a delayed reveal, like how QT does it.
Tarantino answers the rat question at the best possible time. Right when we’re sure we’re going to see the lunatic Mr. Blonde set a cop hostage on fire for shits and grins, the seemingly passed out Mr. Orange shoots him and reveals himself as an undercover cop. It’s a startling moment even on the tenth watch.
How did this heist go wrong? The answer is Mr. Blonde being a lunatic and cutting off a cop’s ear while dancing to “Stuck in the Middle With You.”
Who is the rat? The answer is Mr. Orange laying in a pool of blood, suddenly shooting Mr. Blonde.
That’s exciting shit.
The third major dramatic question — who the fuck are these guys? — is answered less bombastically. We flashback to learning who Mr. White is, who Mr. Blonde is, who Mr. Orange is at different points of the movie when its the most dramatically opportune time to do so.
And these flashbacks and the fallout build up to a new, final dramatic question at the end: will Mr. Orange tell Mr. White that he’s a cop? and if so, how will Mr. White respond?
Structurally, I’d argue it’s impeccable. It displays a kind of showmanship that wouldn’t be available in a more linear approach.
Not all stories are best served being told out of order, obviously. But I think it’s clear that Reservoir Dogs is. And it’s not because of mere cleverness and showing-off. It’s because, through this structuring, Tarantino is able to immediately pose his three main dramatic questions — how did this heist go wrong? who the fuck are these guys? who is the rat? — at the start of his film with genuine urgency.
But perhaps even more importantly, his loose and judicious use of a non-linear structure allows him to move back and forth in time to provide deeply satisfying answers to each of these questions at the most dramatically opportune times in the story.
I would venture this: the best dramatic structure for a script is whatever structure allows you:
to pose the script’s key overarching dramatic questions as early as possible, and in the most urgent and gripping manner as you can manage, and
to answer those urgent, gripping dramatic questions in the most entertaining and/or artful manner you can manage
In some cases, linearity works best for this. A story that’s interested primarily in cause-and-effect, or which operates via suspense, or that offers other narrative complications (such as the rules of some sci fi or fantasy universe) probably should be told linearly.
For instance, my favorite film of all-time is the original The Bad News Bears. It’s the kind of naturalistic film that lives and dies on the subtly changing character dynamics of the team. And that kind of story only really makes sense as a linear film, otherwise it’d be too difficult to track the shifting dynamics and the rising and falling fortunes of the team.
Another hardcore favorite of mine is Robert Altman’s Nashville. And it likewise can only really be properly told in a linear fashion. This is mostly because there’s so many characters and mini-stories to follow and throwing some temporal hijinx into the mix would be too much. (Though I’d argue that Altman and writer Joan Tewkesbury’s multi-faceted ensemble storytelling approach allows Nashville to shift perspectives and dramatic emphases in a way that’s akin to a well-handled non-linear film.)
But some films do benefit from jumping around in time. And sometimes that jump can occur at an unexpected juncture.
For instance, Akira Kurosawa’s Ikiru. It’s a naturalistic film, dramatizing the life of a modest bureacrat who discovers that he’s dying and finds himself suddenly searching for meaning in his life in the little time he has left.
For most of the film, the story follows our protagonist in a pretty traditional manner: we meet him at a point of stasis, then we see as life introduces a dilemma, and then we follow him as complications arrive in its wake.
But then at a key point in the film, well over the halfway point, the story suddenly leaps ahead, after the protagonist’s death. We find ourselves at his wake. It’s disorienting. Upsetting. All of the sudden, a character we’ve grown to love and care for is gone, and we're left with a bunch of lesser individuals discussing and judging him.
But then in a masterstroke, the film reverses gears again as we see our protagonist in his last moments, after we’ve learned that he’s died. And these closing moments are among the most powerful, emotional moments I’ve experienced in any art of any kind. And they would only be made possible in this structural, non-linear arrangement, when we’re most desperate for an answer for the film’s key dramatic question: what makes a life worth living?
There are many elements that make Ikiru one of the greatest films ever made. One of them is that — when necessary — it manipulates time in order to answer its key dramatic question in the most emotionally beautiful way possible.
Wonderful essay. I’d love to see a return of ensemble disaster pics. It’s so fun to imagine yourself as part of the group like in Towering Inferno or Posiden Adventure and wonder how you’d react. I thought The Usual Suspects jumped around in time to great effect as well.
Excellent post, Tony. Will share with my students.